A Modified Mode-Pursuing-Sampling Based Optimization Method for Minimization of Injection Molding Warpage

Yi-Min Deng,Di Zheng,Yong Zhang

DOI: https://doi.org/10.1109/icnc.2009.139

2009-01-01

Abstract:Minimization of the warpage defect is an important task in injection molding optimization. Surrogate modeling technology has been employed to address the problem of intense computation from CAE simulation runs in obtaining the warpage values. This, however, is often faced with a dilemma, in that the accuracy of the surrogate model is dependent on the number of sample points, while more sample points will mean more computation. In this paper, we propose a modified mode-pursuing-sampling (MPS) based optimization method to adaptively develop the sample points during the optimization search process, while at the same time to address the problem of premature convergence incurred from the MPS method. This is achieved by strategies such as, to execute MPS algorithm twice at each generation of the search process, where the latter is used to develop additional sample points outside the local search region; and to apply a global optimization at each generation so as to provide a guiding direction for the search process. A case study of a power outlet cover demonstrates both the effectiveness and the efficiency of the methodology.